Professional Documents
Culture Documents
2 pp. 85-139(55)
9 UK.
10
11 * Corresponding author
12
13 ABSTRACT
14
18 techniques traditionally applied to public health protection are now seeing broader
19 application for asset management, assessing competition risks and potential threats to
20 the security of supplies. Water utility managers have to consider these risks alongside
21 one another and employ a range of techniques and devise business plans that prioritise
23 and management strategies for application in the water utility sector at the strategic,
25
1
1 INTRODUCTION
3 A. Background
5 Providing wholesome, affordable and safe drinking water that has the trust of
6 customers are the overarching goals of the water utility sector. The sector has publicly
7 stated8 that achieving this requires, at a minimum, that water is safe in microbiological
8 and chemical terms; that it is acceptable to consumers in terms of taste, odour and
9 appearance; and that the supply is reliable in terms of quality and quantity. Delivering
15 transition (e.g. Lifton and Smeaton82). In practice, water quality managers and
16 internal audit functions within the sector are working more closely to address issues of
17 business risk and many of the larger international water companies now have group
18 risk managers in place to manage business and consumer risks within a single
21
22 (i) integrated frameworks for the management of internal risks (e.g. from
24 the utility;
25 (ii) the support of Board level, executive management and operational staff as
2
1 (iii) the effective communication of risk and engagement within decision-
5 managing the risks of a commercial water business and the overarching public health
6 goal of the water industry, stated above. Critically in this regard, the transition to an
7 explicit risk management philosophy within the water utility sector is now reflected in
8 recent revisions to the World Health Organisations (WHO) Guidelines for Drinking
10 implementation of water safety plans for water quality management and, within
11 these, the application of risk frameworks and risk tools such as the hazard analysis
12 and critical control points (HACCP)34,64 approach as a basis for prioritising risk
13 management measures within the water supply chain from catchment to tap. The risk
15 a maturing view of risk analysis, shifting from that of a one-off technique to placate
21 making processes (e.g. Pollard et al.121); and technical, relating to the selection and
22 application of risk analysis tools. One of the key difficulties all organisations face in
24 corporate objectives, business plans and operational reality. Here then, we critically
25 review the risk analysis strategies and tools and techniques available for risk analysis
3
1 (strategic), business (programme level) and operational levels in water utilities.
2 Necessarily the discussion requires excursions into the management and technical
14 an agent with the potential to cause either harm and/or benefit (e.g. a
17 likelihood of occurrence);
20
21 The exploration of these facets provides us with an analysis of risk (note that the
22 authors consider the terms risk assessment and risk analysis to be interchangeable).
24 prominent role in many risk analysis strategies. Finally, and in a distinct business
25 context, we consider risk management as the sum of the constituent sets of socio-
26 technical decisions and actions taken by staff to optimise their organisations exposure
27 to risk.
4
1 Risk analysis plays a role alongside other decision tools for risk management.121
2 Detailed risk analysis is not a prerequisite for effective risk management. In many
3 industries there are accepted standards of performance and codes of practice (e.g.
4 engineering standards; accepted best practice; Figure 1) that, if adhered to, provide
7 However, complex, uncertain and novel systems, that deviate from routine operation,
8 may require risk analysis, so as to better understand what drives the risk from or to the
9 plant, process or operation, thereby allowing management measures for the reduction
12 business. This said, risk analysis is, in many respects, a practitioner-driven discipline.
13 Its application within water utilities has its roots firmly in the protection of public
14 health from pathogens afforded by the multiple barrier approach to raw water
15 treatment. Whilst the extension of risk analysis to asset management, water supply
17 and the use of risk-based techniques for optimising treatment plant performance, on-
18 site energy use, maintenance programmes and compliance monitoring regimes can
19 inadvertently but easily detract from and confuse the principal purpose of the water
20 supply industry to provide wholesome, affordable and safe drinking water that has
21 the trust of customers. In all these applications this goal must remain paramount.
22
23
25
5
1 The organisational hierarchy that exists even within flat organisations requires
2 that risks are actively managed at the strategic, programme and operational levels of
3 an organisation (Figure 2). Typically, there are split accountabilities for these risks
4 such that the chief financial officer / financial director and Board have overall
6 of strategic risks; executive and senior management address programme level risks
7 (e.g. asset management, maintenance planning); and operational (e.g. site) managers
8 bear responsibility for operational risks (e.g. treatment plant performance).121 A range
9 of strategies exist for assessing and managing these risks in a business context. The
10 focus in this review is sector-specific, addressing process risk analysis (i.e. risks at
11 the operational and programme level), but in establishing a business-wide context for
12 this activity we also draw upon the experiences of organisations assessing risk at the
13 strategic level.
14
15
17
19 of process reliability, utility managers are increasingly concerned with managing the
22 staff retention, the long-term viability of investment decisions, and the management of
23 external interfaces with regulators and competing utilities. Risk analysis tools are
25
26 A. Regulatory risk
27
6
1 Throughout the 20th Century, the central role of water quality to the protection
3 the public sector.138 Regulation was historically self-imposed and limited in scope,
4 and, by extension, posed relatively low risk to municipalities and utilities (in terms of
5 both the likelihood of non-compliance and the associated penalties). In contrast, more
6 recent (since the 1980s) regulatory pressures and drives to impose market discipline
10 regulation.
12 arises from the nature of the regulatory rules and practices, with rules determining the
14 interpretation the regulators and others (particularly government) place on the rules.
15 Kilpatrick and Lapsey75 consider regulatory risk as the uncertain impact of regulatory
21 the risk of political interference in a regulatory regime. For example, in England and
22 Wales, the economic regulator (the Office of Water Services; Ofwat), acting in the
23 public interest, is vested with a high degree of autonomy from central government,
24 ensuring that the regulatory process is not subject to direct political interference. In
25 contrast, in South Australia (SA), the state government directly controls the tariff
26 setting process, and as the dividend from SA Water is a significant contributor to the
7
1 state budget, there is the danger that political considerations, as well as commercial
4 the importance of set duties and objectives and to determine how best to accomplish
5 them.114,75 In the UK, Ofwats Director General is free to identify and change the
6 importance attached to set objectives within the regulatory system, within broadly
7 defined constraints.114 Arguably, the greater the discretion afforded to the regulator,
8 the greater the uncertainty related to future regulatory decisions. Ofwats regulatory
10 practice, utilities are fully engaged in regulatory decision-making, with avenues for
11 consultation and appeal established should companies wish to challenge the outcome.
12 Similarly, the New Jersey Board of Public Utilities, which regulates all investor-
13 owned utilities in the State, publishes reports on its activities and is transparent and
16 municipalities and the municipalities regulate and manage the water supply based on
17 European, national, state and municipal legislation. Though many are satisfied with
18 these relationships, there has been criticism in a recent report,7 where regulatory
19 decisions were viewed as being taken in a closed fashion with little clear
20 accountability.
21 The nature of a regulatory system (i.e. its objectives and the systems in place for
22 their achievement) represents a core strategic risk for water utilities. For example, in
24 secure water supply,7 which, whilst introducing inherent operational risks, does not
8
1 competition within the sector, an objective that introduces utilities to a range of
4 ex-post analyses of the relationship between utility share price volatility and the
6 (market) risk, using a variable (which measures the variability in returns of a stock
7 relative to the variability of the broader market), of UK water utilities over time,
8 examining the extent to which observed variations were associated with the regulatory
9 process. A key finding was the surge in the markets assessment of the systemic risk
10 to the industry accompanying the surprise result of the 1992 general election. The
11 authors analysis illustrates the influence of politics in even the most independent of
12 regulatory systems. Similarly, Morana and Sawkins98 modelled the London stock
13 markets response to the 1994 periodic review of water price setting in the England
14 and Wales utility sector, finding a significant reduction in share price volatility, which
17 Ideally, for the active management of regulatory risk, analyses should extend to
18 ex-ante treatments of risk. This is of particular relevance to the modern water utility
19 sector, where widespread structural reforms are requiring utilities to operate under
21 market, there has been no historical evolution and the participants, including the
22 regulatory institutions, have a limited understanding of how it will operate in the short
23 term and evolve in the future.78 In such situations, analytical models may offer value
24 in alerting utilities to unintended consequences of their actions that may trigger the
25 regulator into reaction.78 Larsen and Bunn78 argue that system dynamics, which
9
1 uncertain and subjective nature of assumptions inherent to strategic analysis. To
5 describing the main feedbacks involved in the exercise of latent market power.
7 exercise this power and the risk of regulatory scrutiny. The authors concluded that
9 system sensitivity, could provide utility managers with a priori insights into
11 corporate strategies.
12
13
14 B. Competition risks
15
16 Comparative competition
17
19 governments and their regulators have sought to expand the role of sectoral
20 competition. None more prominently perhaps than in the UK, where the concept of
23 proposed a regime in which the price (or financial rewards) received by a regulated
25 regulation), but rather on the costs of identical firms operating within the same
26 sector. Shleifer reasoned that by breaking the dependence between the price a firm
10
1 received and its own costs, and ensuring that the rewards for a given firm depended
2 on its standing vis--vis a shadow firm (a weighted average of other firms operating
3 within the sector an idealised benchmark), each firm would be forced to compete
4 with its shadow, providing incentives for cost efficiency (widely perceived as lacking
5 from rate of return regulation). In practice, the inherent risks of this competition by
6 proxy pale in comparison to those found in fully liberalised markets because market
8 Techniques for evaluating the explicit risks posed by competitors have been
10 competitor analysis, with its potential to reduce the uncertainty of the price review
12 helped by the tendency for regulatory bodies to disclose company performance data in
16 assessing the implications of competitors strategies on both the sector and their own
17 utility.38,133
18
20
26 mechanism for utilities, as inefficient firms would be reflected in their share price and
11
1 be vulnerable to take-over, in addition to facing higher costs of capital. Although
6 utilities.
8 under the yardstick system assists predators in identifying and assessing potential
10 surmised that profitability, size, leverage, and dividends were negatively correlated
11 with the risk of being acquired. In contrast, profitability and liquidity were positively
12 correlated with the probability of a firm acquiring, with leverage and dividends
13 negatively so. In light of this information, the dynamic risk of take-over can be
14 tracked both in real-time (e.g. with respect to the transfer of, for example, more than
15 5% of firm stocks to a potential acquirer and, in the US, the filing of 13D statements
16 indicating investor intent) and pro-actively (by screening the external environment
17 for trends and potential hostile bodies). Of further interest to corporate strategists,
19 strategy to reduce the risk of take-over. The researchers concluded this strategy
20 allows firms to grow quickly, thus protecting them from subsequent take-over. For
23
25
12
1 Another means of fostering competition is to encourage the private sector
2 (perhaps along with the incumbent public utility) to bid competitively for a
3 concession, lease, tender, or management contract.27 The two key vehicles for doing
4 so are franchising and, more conservatively, contracting out (not involving the
6 including: build, operate and transfer (BOT) arrangements; finance, operate and own
7 (concession); and operate and provide working capital (affermage). The inherent
8 complexity of many of these arrangements, the generally low equity in the project
9 vehicle,55 and the often significant investment obligations required of the sponsor,
11 The project and financial risks associated with public-private partnerships have
12 been reviewed by Grimsey and Lewis.55 Using the financing of Stirling Water, a
16 flows from each of the senior lenders to this joint public-private venture. From the
17 procurers perspective, project risks (e.g. delays and claims) are valued and
18 incorporated within the NPV calculation, whilst the impact of financial risks (inflation
19 and interest rate changes) are evaluated through sensitivity analysis. From the
20 sponsors perspective, risk analysis centres on simulating the effect of the underlying
21 variables (e.g. operating performance) upon the equity return. Ranasinghe128 uses
22 water supply projects in Sri Lanka to outline a methodology based on financial risk
23 analysis that a government or public utility can use to assess the viability of private
13
1
5 utility services has been to separate the monopolistic component of the industry and
6 regulate it, and to encourage competition in all other areas, e.g. the UK model of
7 separating the gas, electricity and railway networks (monopolistic) from the supply of
9 promoted by the World Bank,172 has not been widely adopted in the water sector, the
11 has led the way in adopting alternative approaches to facilitate product market
12 competition. This can be traced back to the 1991 Water Industry Act, which
13 introduced the concept of Inset appointments, whereby a utility can apply for an
14 appointment to provide water to a large customer located within the statutory area of
16 Sawkins136 reports that the first Inset appointment was granted in May 1997, when
17 Anglian replaced Essex and Suffolk Water (ESW) as the supplier to Buxted Chickens
18 Ltd. Company licences were altered and a new pipe constructed linking the site with
20 In practice, various restrictions, recently eased, have meant that this form of
21 competition has been slow to develop.136 Similarly, although the 1992 Competition
22 and Service Act allows for cross-border competition, the costs are prohibitive in the
23 majority of cases. Perhaps the most significant recent development has been the
24 introduction of the 1998 Competition Act, which created the possibility for common
25 carriage agreements, or network sharing, in the water industry. Here, the shared use
14
1 within the incumbents area. To aid this, Ofwat now requires that all water utilities
2 publish Access Codes that set out their terms and conditions for common carriage,
3 and has published guidance on this procedure.110 Hern65 reports that under the Act,
4 utilities risk infringement if they refuse access to any parts of their infrastructure
7 have resulted to date, the threat alone acts as a catalyst for performance
8 improvements.
9 The authors were unable to uncover literature quantitatively addressing the risks
10 of product market competition within the water utility sector, a reflection of its
15 with a critical appraisal of self-performance and room for improvement, often provide
18 analysis, and in the absence of a relevant body of literature, the authors suggest
21
23
24 Our discussion thus far has focused on the strategic approaches to risk
25 management within the sector. The pressures described are having important impacts
15
1 increasingly demanding political and consumer environment, and more stringent
4 finding new ways to address problems, and revamping traditional business models
6 engineering (BPR) initiatives which range from the redesign of existing processes
8 new corporate vision require the commitment of substantial resources and often
9 constitute a lasting legacy. If we define the risk of a project as the deviation in results
10 from the established goals, then there is substantial empirical evidence marking BPR
11 as a high risk endeavour. Many, if not most re-engineering efforts ultimately fail
12 (see Crowe et al;28 Remenyi and Heafield130). Of particular relevance is the work of
14 industry suggests that re-engineering efforts, whilst often effective, produce highly
15 variable outcomes. On account of this, project risk analysis should be an integral part
17 Clemons23 considers the core determinants of the risk profiles associated with
18 large scale BPR efforts to be: (a) functionality risk the risk of making inadequate or
19 incorrect changes to systems or processes; and (b) political risk the risk that the
20 organisation will not complete the project, either because of significant internal
21 resistance to the proposed changes or due to a more gradual loss of will. Clemons
23 a means for assessing and subsequently managing the risks associated with re-
24 engineering efforts. Rather than determining a single correct view of the future with
25 its implicit single response, scenario planning acknowledges the key sources of
16
1 strategic responses for exploration. Clemons argues that its use is suited to the
3 futures and strategies, reduces functionality risk and helps ensure the need for change
4 is internally addressed and accepted, thus reducing political risk. Scenario planning
5 has been embraced by the majority of UK water utilities.120 A 2001 study120 explicitly
6 linked the tools use with improved financial performance on the part of utilities,
7 although notably the authors suggest that scenario planning may implicitly encourage
10 tool for estimating the risk of failure of companies about to undertake re-engineering
14 systems) and failure (middle management fear of losing authority; fear of job loss;
16 Raw data is extracted by questionnaire (e.g. do managers usually share vision and
18 leadership style), and refined via fuzzy mathematics. Crowe et al.s model is intended
22 Heafield130 outline a methodology for evaluating the key risk issues relating to re-
23 engineering efforts. The methodology centres on a risk matrix (Table 2) that groups a
24 variety of potential BPR risks under the categories of business risk, financial risk,
26 weight and rank what they consider to be the ten factors most pertinent to their
17
1 proposed re-engineering efforts. The framework represents a succinct method for
2 appraising and comparing the risks associated with BPR strategies. A perceived
3 failure of much of the BPR literature is the limited emphasis placed on the risks
5 efforts.
7 Technological risk
10 concern of strategic planners and policy makers within the water industry. The
11 introduction of novel technology poses risks due to the inherent difficulty of preparing
13 components and processes; and the long development cycles required for changes in
14 regulations and consumer demands.26 This has led many researchers to advocate the
16 technologies (e.g. Colmer et al.26; Fitzpatrick47). This is highly relevant to the water
19 advanced oxidation and other novel physical and mechanical technologies; the broad
20 use of membrane systems to desalinise seawater for human consumption; and the
22 McGaughey et al.92 describe a framework for viewing and comparing the risks
24 proposed projects are assessed, through value chain analysis, in terms of their
25 potential positive and negative outcomes these are then mapped onto a speculative
18
1 stages of planning, specific threats and opportunities associated with the project are
3 Hartmann and Lakatos62 drew on case studies monitoring the pace and quality of
5 algorithm characterising the risk of each technology problem (Figure 3). The authors
6 suggest that its use can aid in the refinement of technology development and
16
19 tool for minimising the risk associated with introducing new technologies (i.e. beyond
21
19
1 - failure modes identified
5 mathematical simulation
12 impact study
13 Contingency planning
16
18 profiles are constructed displaying the relative importance of identified threats and
21 successfully implement the technology. The authors aim was to provide an analytical
22 basis upon which strategies may be developed for the introduction of new
23 technologies.
24
25
20
1 D. Outsourcing risks
3 Our discussion of risk analysis strategies moves to one of the key features of the
5 management in recent years has been the growth in outsourcing, defined as the
11 customer care and billing.116 A widely held view is that the potential for outsourcing
12 is far from exhausted. A holistic approach to risk being promoted in this review
15 fall within the remit of corporate risk management. That is, outsourcing alters the
16 boundaries of the firm, and the scope of risk analysis and risk management
18 Risks are inherent in the process of outsourcing, from the decision to outsource,
19 to the management of agreed contracts. Received wisdom has been that companies
20 should focus on core competencies and outsource the remaining parts of the
21 business (although the validity of this distinction has been questioned of late, notably
22 by Heikkil and Cordon63). The core risks discussed in the literature relating to
23 decisions over what to outsource and who to outsource to include: the loss of key
24 capabilities, developing dependence on the vendor, and risks linked to the service
21
1 available for this purpose. Lonsdales86 decision tree for outsourcing provides a
3 analysing market opportunities for outsourcing the remaining parts of the business.
4 The framework seeks to ensure managers retain those resources responsible for
9 the transpiration of failures with in-bound goods and services.176 Core categories of
10 supply risk discussed in the literature include: the financial stability of the supplier;
11 cost fluctuations; capacity constraints of the market and specific suppliers; variations
14 managing supply risk are the active management of risk interfaces with the intention
15 of reducing vendor failures,176 and the construction of barriers (e.g. safety stock,
21
22
23 E. Employee retention
24
26 utility management. The recent emphasis on people as the resource, along with the
22
1 external realities of an increasingly dynamic and pressurised labour market, have led
2 to the sector embracing employee retention as a critical risk issue particularly in the
3 technically specialised areas of the water business. This focus is exemplified in recent
6 the information age), and a recent (2001) policy statement from AWWA calling on
10 Empirical evidence suggests that ensuring a cultural match between employees and
11 the organisation plays a critical role in reducing staff turnover.141 The tool applied by
12 Sheridan141 to measure culture (beliefs and values) was the Organisational Culture
14 candidates by encouraging them to sort value statements on: norms regarding the
16 regarding individual actions. Utilising the OCP as a part of the recruitment process
17 could provide utilities with a proactive tool for minimising staff turnover, by filtering
18 those most likely to leave the organisation early from the selection process.
21 This philosophy is mirrored in the work of McNally,94 who promotes the use of
23 ensure good fits of personality and work ethic. As Denton33 notes, whilst good
23
1 identify employees at risk of leaving. Such a system requires the collection and
5 local economic trends, head-hunter activity, and, crucially, the exit interview may be
7 system would provide utilities with comprehensive data on who is leaving and why,
9 recurrent theme of the retention literature is that incentives (e.g. salaries and benefits)
10 alone are not enough for achieving high levels of retention, the contention being that
11 retention is related more closely to employee development and intrinsic benefits such
15 Brueck18 reports that water utilities spend as little as 1% or less of their labour budget
19 which is essentially a checklist addressing the core issues influencing turnover (e.g.
22 throughout the literature, leading to the implicit assumption that organisations should
23 pull out all the stops to minimise defection rates. However, as Sigler144 and
24 Mowday100 contend, the costs of reducing retention may, in some cases, exceed the
24
1 costs and benefits of implementing retention strategies; the cost-benefit analysis
8 strategic investments as those resource allocations that will yield substantial advances
12 turn a key driver of its overall performance.87 Valuation methodologies range from
13 the formal (comprising an appraisal model and a supporting theory) to the informal
14 (based on heuristics).87 However, since the 1970s there has been a trend towards
15 applying valuation methods that are more formal, explicit, and institutionalised.87 The
16 most widely adopted framework is the Net Present Value (NPV) model, which
17 estimates value by capitalising (discounting) future streams of cash flow that the
18 investor expects to receive from an asset. The capitalisation rate is the minimum
20 comprised of two components, the risk-free rate of return (accounting for the time
21 value of money) and the risk premium (the additional compensation demanded by
22 investors for assuming risk). Although issues have been raised regarding the
24 specifically relating to the long lifespans of many capital projects and the fact that
25 they often do not generate revenues in the traditional sense (e.g. Tebbutt et al.155),
25
1 focuses on three distinct investment problems: valuation of assets-in-place; valuation
4 Assets-in-place
6 The most basic valuation problem is valuing assets-in-place, i.e. the valuation of
7 an ongoing business or some part of one, for the purposes of informing decisions
10 capitalisation rate) are suited.87 In brief, the established DCF methodologies include
11 the weighted-average cost of capital (WACC),97 the capital asset pricing model
12 (CAPM)140 and the adjusted present value (APV).102 The WACC, which establishes
13 the risk premium on the basis of the cost of capital financing the investment,
15 academic circles (e.g. Luehrman;87 Gregory54). The fundamental idea behind CAPM
17 APV seeks to unbundle the various components of value (i.e. cash flows), analyse
18 them separately, and then add up the present values. For a fuller discussion of these
19 and other DCF techniques see e.g. Modigliani and Miller,97 Sharpe,140 Myers,102 Berry
26 principle being that investment forecasts are, by definition, uncertain. Reflecting this
26
1 uncertainty in model outputs lends some assurance to the decision-makers that the
2 available information has been used with maximum efficiency.66 This is reflected in
6 examine the influence of changes in key underlying variables on forecast cash flows,
7 and the probability that project NPV will fall below zero. Incorporating these
8 principles, Barriex et al.10 describe the application of the NPV framework to the
13 Through stochastic simulation and sensitivity analysis of forecast financial returns, the
17 adjusting for the risk inherent to acquisition / diversification appraisals (using internal
18 rate of return (IRR), an appraisal framework similar to NPV). Accounting for the
20 methodology for integrating expected financial and operational synergies (e.g. derived
21 from financial and operating economies, or the pooling of functional areas) within the
22 analysis. However, through applying a risk premium to projected cash flows (which
23 by definition accounts for the increased returns investors demand for variable cash
24 flows) and undertaking simulation of the variables influencing future cash flows (thus
26 counting for risk, introducing a bias against investment decisions. This criticism is
27
1 supported in the work of Burchett and Tummala,21 who apply Monte Carlo simulation
4 variables captures all potential risks relating to the investment, thus negating the
7 analysis is desirable, problems exist. As Songer et al.148 assert, the failure to identify
8 all significant risks (i.e. to apply appropriate probability distributions to all relevant
9 underlying variables) quickly undermines model validity and output. A further pitfall
11 proposed plant investment, found that the choice of frequency distribution chosen
12 (often arbitrarily) for the independent variables can have a marked effect on the
13 process outcome. These are important observations in that they highlight the biases
14 inherent to all risk models, reminding of us of the need to use risk analysis output
16 In financial circles, recent times have seen an increasing adoption of tools that
18 investments (portfolio) as well as for each individual project. This trend extends
19 beyond the financial sphere, as is illustrated in the work of Rothstein and Kiyosaki,134
22 representing a balanced array of investments that mitigate uncertainties and that are
26
28
1
2 Opportunities
5 second type of valuation problem the valuation of opportunities (i.e. possible future
7 distinction is that with the former, the decision to invest may be deferred. In
8 opportunity valuation, risk matters in two ways: the risk of the investment, and the
9 risk that circumstances will change before a decision has to be made such
10 contingencies are not well handled by the traditional DCF approach.87 Luerhman87
11 states that a common approach in the valuation of opportunities is simply not to value
12 them formally until they mature to the point where a decision can no longer be
13 deferred, where they can then be valued, in effect, as assets-in-place. Critics have
14 decried this practice, on the premise that it leads companies to undervalue the future
18 supplement, not a replacement, for the valuation method for in place assets.
19
20 Joint-ventures
21
24 in the water industry, where recent years have seen a proliferation in public / private
25 partnerships. In such cases, where ownership is shared with other parties, managers
29
1 need to understand both the value of the venture as a whole and the value of their
4 reviewed by Grimsey and Lewis.55 Using the financing of Stirling Water, a Scottish
6 of returns on investment from the perspective of the private (sponsor) and public
9 obvious gap between theory and practice. Much of the highly theorized investment
10 literature does not reflect standard industry practice, particularly that relating to the
12 discrepancy is explained, in part, in that such techniques do not fit naturally into most
14 literature focussing on the practicalities of integrating such tools deep within company
15 structures. To address this issue and as part of the research that has informed this
18
19
21
23 water utility sector. The revised WHO guidelines167 are promoting the implementation
24 of water safety plans for water quality management from catchment management,
25 through process control, distribution and on to the tap.160 Application of risk analysis
26 to these aspects of the water supply chain extends to programmes of work as well as
30
1 individual plant operations. A discussion of the latter, operational risk analysis
2 follows, but here we are concerned with the analysis of risks associated with
4 management and maintenance planning. Here, managers are responsible for the
6 regions. They are concerned with: evaluating the risks posed by a similar hazard at a
7 variety of locations (e.g. mains bursts, network intrusion in asset management, for
9 relation to the water supply-demand equilibrium; and the wide variety of risks
12
13
14 A. Asset management
15
17 infrastructure capital assets to optimise the total cost of owning and operating them
18 while delivering the service levels customers desire. Managing risk in the face of
19 limited resources has long been an implicit component of asset management. Within
20 the UK, pressure from the economic regulator has ensured that the explicit
21 incorporation of risk analysis into asset management programmes has taken on added
23
25 cost in terms of both capital maintenance and operating expenditure, recognising the
31
1 trade off between cost and risk, whilst ensuring compliance with statutory duties
5 self-sufficiency on the part of public and private utilities has created a climate in
8 Senate cites mounting evidence suggest[ing] that the integrity of the nations
11 replacing these assets. The report goes on to explicitly endorse the role of risk
12 analysis in asset management. More than ever, utilities must now seek to balance
16 the process to the component level (e.g. Lifton and Smeaton;82 Booth and Rogers16).
18 recently (2003) reported on the risk to drinking water quality from ageing pipes and
19 process plant across the US with individual city rankings being informed by water
20 quality data, USEPA compliance records and water utility annual reports. Many
21 water companies have in place risk-ranking procedures to evaluate and rank potential
22 risks across a variety of categories, and thus help inform and prioritise risk
26 KANEW, a statistically based survival model which aids the calculation of pipe
32
1 rehabilitation and replacement needs for distribution networks. The identification of
2 specific pipes requiring work is external to the model, with separate approaches for
3 trunk and reticulation mains (the latter generally being run to failure). Critical trunk
6 an overall risk score. This combined risk score is used to identify critical water mains
8 methodology allows Sydney Water to identify and prioritise water mains in need of
11 which integrates data such as pipe age and maintenance history, as a tool for
12 prioritising pipe and water mains for rehabilitation and replacement.163 Utility
13 managers report that this model, in combination with wider asset management
14 practices, has helped reduce the frequency of water mains breaks from 26 to 22.7 per
15 hundred miles and the frequency of joint leaks from 8.2 to 5.6 per hundred miles.163
17 likelihood and impact of pipe rupture with reference to such factors as age, material,
18 location and historical cost of repair.163 Drawing upon this analysis, utility officials
19 were able to delineate their pipe network into areas of critical and non-critical risk,
21 this approach, officials believe that they are using staff resources more efficiently and
22 that, over time, the programme will lead to a reduction in maintenance costs.163
23 Kent et al.73 describe how risk analysis informs the prioritisation of investment
24 strategies for trunk main maintenance at Dwr Cymru Welsh Water. The methodology
26 which are stored on the companys WAM (Water Asset Management) database.
33
1 STRUMAP, a software-based mapping system, allows clustered failures to be
2 considered separately from random bursts, a task performed as the former are
4 where a cluster is identified, specific failure rates are derived. For random bursts,
5 failure data is separated according to pipe material and diameter, with failure
8 taking into account service reservoir storage. Failure likelihood and consequence are
9 then combined to derive an overall severity score, which in turn informs the
12 aimed at integrating the most promising breakage analysis models into one discrete
19 weighting the historical frequency of piping failure with respect to dam zoning, filters,
20 dam age, core soil types, compaction, foundation geology, dam performance, and
22 risk for more detailed analysis, and is further offered as a check on traditional event-
25 engineering technique that tabulates failure modes of equipment and their effects on a
26 system1 (Table 4). The failure mode describes how equipment fails (open, closed, on,
34
1 off, leaks, etc.). The failure effect is determined by the systems response to the
4 Lifton and Smeaton82 detail how Scottish Water apply source-to-tap FMECA
5 studies across their water supply systems as part of their asset management toolkit.
6 This allows priority risks to be identified and subsequently compared across the utility
7 portfolio (e.g. various mains, raw and treated reservoirs, treatment works etc.) in order
10 model. Of particular interest is their description of the asset risk and criticality
11 scoring system implemented at Scottish Water. The system is designed to assess the
12 relative total business impact of asset failures across the company by reference to a
15 this scoring system guides the prioritisation of reliability studies at the operational
18 assessing their likelihoods,53 logic models (visual risk schematics, e.g. reliability
19 block diagrams, fault tree analysis (FTA) and event tree analysis (ETA), see Figures 4
20 and 5) have found application in support of asset management. Parr and Cullen,117
22 analysis, illustrate how such an approach can inform the prioritisation of expenditure
24 promote the use of logic diagrams in aiding the development of risk-based strategies
25 for maintaining asset security. The authors model the interactions leading to failure
26 for each class of aqueduct structure. To this, historic data, or where data is deficient,
35
1 engineering judgement, are applied in order to derive failure probabilities. A
2 cautionary note is sounded by Latiffe,79 who contends that risk analysis, specifically
3 logic modelling, is not yet effective in modelling dam failure. The author cites
5 drawback.
10 asset management. At its most basic level, GIS allow utilities to convert data
11 ordinarily displayed on paper maps into one single, easily accessible digital format,
12 representing an excellent method for storing and collating data for future use.49 The
13 level of detail (i.e. the layers of spatial data) contained within such systems varies
14 widely. Kaufman and Wurtz72 describe the evolution of a GIS for a small utility
16 records and failure and maintenance data is collated within the system, supporting the
18 system took only three months and less than $3,000 to develop. Similarly, Booth and
20 management decision support system can allow for the visual tracking of
24 rarely utilise the capabilities of GIS to spatially analyse data in the classical sense.49
36
1 providing a basis for the identification of network areas most at risk from external
2 corrosion. Such an approach may allow utility managers to better focus rehabilitation
3 efforts through having a more complete understanding of the causative factors behind
4 water main deterioration. Of further interest is the work of Ta,153 who describes the
5 application of a probability model for burst risk studies of water mains. Contributing
6 factors (e.g. pipe number density, pipe age, material and diameter, soil corrosivity,
7 etc.) are represented as GIS data layers and correlated with past failure data in order to
8 deduce burst probability scores for each water main. The tool, developed for Thames
9 Water Utilities Ltd. (UK), is not intended to predict the likelihood of pipe bursts,
10 rather to aid utilities in sourcing the origin of an area burst (i.e. following a pipe burst
11 in the area, the value of probability evaluated for a particular pipe section would
12 indicate the likelihood that the burst actually occurred at that section).
13 While GIS represent powerful tools for spatial data analysis, their inherent
14 capabilities for complex and dynamic analysis are limited.152,43 In contrast, traditional
15 simulation models are powerful tools for complex and dynamic situations, but often
16 lack the intuitive visualisation and spatial-analysis functions that GIS offers.152,43
18 Buchberger83 describe the integration of hydraulic modelling within a GIS for the
20 methodology enables the synthesis of multiple risk factors describing the three key
26 may also be applied in a reliability context at the design stage. Similarly, Besner et
37
1 al.13 illustrate via case study how the coupling of a GIS containing structural,
2 operational and water quality parameters with simulation model EPANET facilitates
3 the identification of key factors responsible for water degradation in the distribution
4 network. Through identifying network areas presenting the greatest risk, this
8 B. Catchment management
13 of water quality problems and their integrated solution.48 An outcome of this is that
15 increasingly subject to formal risk assessment and can be expected as part of routine
17 hazards within a watershed, by reference to the driving forces (e.g population growth),
19 eutrophication) and policy response (discharge control) is being adopted under the
21 an actual or potential threat to the quality of water bodies in river basin districts are
23 targeted at raising the overall ecological status of the watershed within statutory
38
1 likelihood of severe impacts being realised. Southern Water (UK) adopt a semi-
4 methodology consists of ranking source waters across ten risk categories (e.g. land
5 use) using pre-determined scoring hierarchies (e.g. occasional livestock grazing: 2),
6 before combining these category rankings into an overall weighted risk score.
7 Through this approach the utility identifies those sources deemed to be at significant
8 risk of oocyst contamination, and which therefore require continuous monitoring (in
10 Given the improved capabilities and functionality of modern GIS and their
11 inherent ability to map and analyse data that is spatially variable in nature, many
14 essentially combine the attributes of two or more data layers across geographic space)
15 in the identification and mapping of areas critical to catchment water quality. These
17 those spatial attributes considered to play a significant role in pollutant transport (e.g.
20 targeted towards specific hazards (e.g. animal feeding operations) or pollutants (e.g.
22 Risk-ranking methods are applied to help target more detailed analysis towards
24 specifically monitoring programmes. Of course, the potential exists that as the costs
25 of planned monitoring decrease on the one hand, the risks may increase on the other.
26 When designed well, piloted and implemented with feedback, risk-based resourcing
39
1 strategies (Figure 6) can provide a sound basis for distinguishing greater risks from
2 lesser ones, and for investing resources in risk management that are proportional to
6 themselves incur significant risk unless the consequences of resource trade-offs are
10 the Saskatchewan drinking water programme and, partly in response to budget cuts in
11 the mid 1990s, drastically reduced the already limited field inspection and
14 Treasury Board in 2000/01 and justified as being risk-based. The subsequent North
16 community plus a large number of visitors from three other provinces, led to a public
17 inquiry into the outbreak and the provincial drinking water regulatory system. Justice
18 Laing76 concluded in his Inquiry report: that the current risk-based model employed
19 by SERM since 1996 is arrived at on the basis of economics, and has nothing to do
20 with how best to safeguard the health of the population, all of whom consume water.
21 The example aptly illustrates the inappropriate use of risk analysis as a justification
22 for the removal of processes critical to public health protection. Tensions that arise
23 between those seeking economic efficiencies and preservation of the principal goal of
24 providing safe drinking water are often played out in the conflicting expectations and
26 consequences of stripping away levels of safety, precaution and protection using risk
40
1 analysis as a justification can be to render the system as a whole less safe, more
4 scrutiny as to the balance between risk and the full cost of implementing these
5 programmes.
7 model-based approaches. Water quality and flow / transport models represent core
8 tools for this purpose, due to their combined ability to model the dispersal of
9 pollutants and predict the resultant deterioration of water quality. Aside from the
11 issues, the core benefits of model-based analysis stems from their ability to test
12 management scenarios (through e.g. sensitivity and scenario analysis), thus enabling
13 informed decisions on how best to manage the resource. A range of models are
14 available that apply to catchment risk analysis, from micro to landscape scales, from
16 Common practices of hydrological and water quality modelling have been based
18 prediction uncertainty.2 Determinism has been embraced by many risk analysts, for
19 example, Gndz et al.56 describe the use of the combined hydrodynamic and water
21 patterns under different pollution loads. The tool is intended to aid management in
24 other researchers (e.g. Cole et al.25). The limitations of determinism in risk analysis,
25 discussed earlier, are particularly relevant in the context of hydrological and water
26 quality modelling, considering the often scarce or incomplete data available.89 This
41
1 uncertainty takes on particular importance from the utility standpoint, as their
2 assessments of catchment water quality are performed with regard to set regulatory
3 standards. To illustrate this point, the uncertainties inherent in flow and contaminant
4 transport modelling (from e.g. spatial variability, data scarcity, model imperfections)
5 imply that there will always be a risk of exceeding a given standard at some point
6 over space or time following a pollution event, regardless of the estimated single-
9 uncertainties in catchment level risk modelling. There exist two dominant approaches
10 towards this task: stochastic modelling; and deterministic modelling allied with
11 uncertainty analysis of the output. Adopting the former approach, Andersson and
12 Destouni2 outline the application of stochastic transport modelling to quantify the risk
15 required to attain an acceptable risk level. Halfacree,61 for example, describes the
17 pollution risks to water bodies. The main elements are an aquatic dispersion model;
19 facility. The model has a deterministic mode used to screen out low risk sites, and a
20 probabilistic mode for more detailed analysis of high risk sites. The output results
25 equations is often impractical for complex problems.81 This explains, in part, the
42
1 preference for deterministic approaches to water quality / hydrological modelling,
4 obtaining a given output value when uncertainties on input variables and parameters
5 are known.89 Liou and Yeh84 outline the use of a groundwater transport model in
7 upper limit (e.g. regulatory standard). The analytical uncertainty of the predicted
9 prior to the application of Monte Carlo simulation in order to compute the mean risk
11 the forms of uncertainty in water quality modelling and the techniques for their
19 relation to this, Latinopoulos et al, through coupling stochastic flow and transport
22 failure in terms of regulatory fines and the need to import / develop alternative
23 supplies).
26 al.;171 Rautman and Istok129). These kriging methods essentially a form of least
43
1 squares linear regression focus on providing an estimate of a spatially distributed
3 limited set of sample values taken from surrounding locations.129 As such, they are
7 mimic the random variability of the parameter(s) of concern are produced.171 Various
8 authors118,171,129 have illustrated how such an approach may answer the following
9 questions: what is the probability that contaminant levels exceed regulatory standards;
10 where are the compliance boundaries (and what is the associated level of confidence);
11 and how much contaminant is present (and hence, how much must be removed)?
12 Although the principles of geostatistical simulation are well established, the technique
16 ability to quantify risk over space and time is limited. To counter this, researchers
17 have sought to integrate these systems with simulation models. Feijtel et al.44
18 illustrate that the embedding of chemical fate prediction models within a GIS allows
21 approaches are adopted by Dabrowski et al.30 and Verro et al.164 to assess surface
23
24
25 C. Network analysis
26
44
1 A water distribution system may be viewed as an interconnected collection of
2 sources, pipes, and hydraulic control elements (e.g. pumps, valves, regulators, tanks)
5 layout, may be described mathematically.112 This description forms the basis of water
7 water industry for many years, particularly to inform the development of operational
8 strategies.154,17 Water utilities routinely apply network analysis in order to assess their
10 demands (i.e. network reliability). Best practice utilities extend their analysis
13 demand balance. The standard Scottish Water methodology of yield assessment uses
15 simulation model, is used to derive historic inflow series, based on historic rainfall,
17 abstractions). AQUATOR, a water resource system model, uses the output from
19 flows. The model assists Scottish Water in understanding the level of supply
20 availability risk in the current system and in determining the impact of prospective
22 Stevens and Lloyd150 describe the application of the resource modelling package
23 WRAPsim, with reference to the Yorkshire Water (UK) Grid. The model contains
24 over 1200 components including all river and reservoir sources, boreholes, water
26 conjunctive use of Yorkshire Waters sources over a given time period, model output
45
1 provides the decision-maker with an accurate assessment of the behaviour of each
2 source, its ability to meet demand, and the frequency of restrictions that would need to
3 be imposed. Further insights are gleaned through the application of scenario analysis,
4 wherein the supply-demand balance for each zone under variable scenarios (e.g.
5 average year, dry year, peak week, etc.) allows an assessment of security of supply
6 over a range of timescales and operating conditions. The authors report that
8 resources, and to rebalance stocks, has significantly increased the yield and reliability
10 Stahl and Elliott149 discuss Essex and Suffolk Water (ESW)s use of the risk-
11 based resource planning and operational support model DROP (Drought Reliable
13 utilitys specific technical requirements. The model has been applied in a variety of
15 operational strategies. The authors state that DROP has enabled ESW to improve
20 adequately address whether the system is sufficiently reliable, as this requires the
23 power and operability, Ostfeld113 has developed a methodology for the explicit
25 and measured as the probability of zero annual shortfalls. The methodology, whose
26 development was funded with the intention of practical application by the Israeli
46
1 Water Commission, is comprised of two interconnected stages: (i) analysis of the
6 the need to take account of both the frequency and severity of modelled failures, and
10 Adopting this paradigm, Zongxue et al.174 describe the coupling of a risk model
13 approach aids the identification of operational strategies of minimum risk under given
15 Supply System, Japan (see also Jinno et al.71). Similar methodologies are described
16 by Wang et al.;165 Merabtene et al.;96 and Andreu et al.3 though supplemented with
22 policies; (c) assist in the design phase of distribution networks; and (d) inform the
24
25
47
1 D. Vulnerability assessments
7 The role of formal risk analysis in emergency planning, long restricted to drought
8 management, is now being widely adopted to address security risks. This is largely in
9 response to the events of September 11th, 2001. In relation to this, a methodology for
13 vulnerabilities and to develop measures to reduce the risks and mitigate the
15 steps:147
16
18 surveying all security and monitoring features (e.g. how quickly could it
21 stop undesired events (e.g. security in place, length of storage time); and
22 3) measure the capacity of private guard forces and local, state and federal
24
48
1 Perhaps a more pragmatic approach, particularly for smaller utilities, is found in
3 Association.107
8 Our review now progresses to the analysis of individual plant. Operational risk
9 managers are responsible for the risks associated with specific operations at plant
10 level for example, the risk of failure of a device or process component, or the risk of
11 exceeding a particular water quality standard and they are increasingly responsible for
12 the health and safety of plant operatives. Analysis at this level is largely concerned
13 with the classic risk analysis methodologies developed and established within other
14 process industries, most notably the oil and chemical sectors (Table 5).
15
16
18
19 Here, we are primarily concerned with the risk posed by specific contaminants
20 at the plant and distribution system level, particularly relating to the hazards posed to
21 human health and the related risk of exceeding regulatory standards. The multiple
22 barrier approach to water treatment has been the central tenet of modern water
23 treatment systems and relies upon the use of in-series water treatment processes to
24 remove hazardous agents from the public water supply. Failure or inadequacy of the
49
1 public health. The underlying causes may include source contamination, human error,
3 immediate, there is very little time if any to reduce exposure because of the lag in
4 securing monitoring data and the impacts can affect a large number of people
5 simultaneously.121 Beyond the paramount impacts on public health through the direct
7 invariably ensue. The financial costs to the community of the fatal Walkerton
8 outbreak for example, were in excess of Cdn$65 million, with one time costs to
9 Ontario estimated at more than Cdn$100 million.108 Compounding this, the loss of
12 against operating companies. Litigation for civil damages have been prominent
13 features following both the Walkerton outbreak (settled out of court) and the Sydney
17 need for a preventative approach to managing risk and providing safe drinking water,
19 methodologies within the sector, for both chemical and microbiological hazards.5, 59
20 The generic approach is based on the risk assessment framework developed by the
23 effects derived from any particular hazard (e.g. infection, carcinogenicity, etc.)
50
1 dose-response assessment to characterize the relationship between the dose
8 agents and assessment of risk of chemicals.58 Accordingly, the NAS approach has
11 assessment (QMRA). The application of these models has long been the basis for the
13 health risk assessments that have historically informed the guidelines may, however,
15 utilities.121 However, recent work has extended the application of these models to the
17 and Teunis et al.156 describe the application of QMRA in determining the public
18 health risks posed by the presence of microbial contaminants in treated water. The
19 first step in the process is to define the relationship between measured pathogen
20 source levels and the consumed dose (incorporating analytical detection levels,
24 output of the deterministic model to determine the distribution of the daily consumed
25 dose, to which the relevant dose response relationship is applied in order to determine
51
1 the cumulative distribution of the probability of infection. From this, the mean annual
3 relevance in areas, such as the Netherlands, where water supply legislation expresses
4 acceptable health risks in terms of infections per year.95 Of course, core microbial
5 standards generally refer to a maximum level of organisms in the treated water, and so
7 to compliance risk assessment. The approach perhaps has most utility in what-if
8 mode to answer questions such as: what are the public health implications of a
10 process.52
11 Tools are available to assess the risk of exceeding water quality standards
14 compliant drinking water across a range of parameters, taking into account the quality
15 parameters of raw water and the removal efficiencies and reliability of the full set of
18 water treatment. These three pieces of research explicitly consider the performance
20 described in operational QMRA. Not only does this offer a more realistic appraisal of
21 compliance risk, it is in line with recent proposals from regulatory bodies (e.g.
23 formally adopt the multiple barrier approach to risk management to ensure multiple
24 levels of protection are afforded against specific contamination threats (see Rizak et
25 al.131).
52
1 Of course, limitations in resources (human and financial) and in the data to
2 underpin such sophisticated analyses often restrict the practical application of these
3 more advanced methodologies within the sector. A more pragmatic analysis of the
7 risks at a water treatment plant. Risks are scored according to the frequency with
8 which they may occur, the ability to take action to contain the event, and the
10 targeting and prioritisation of remedial actions. Such approaches rely heavily on the
11 experience and judgement of the assessment team, and depending on the level of
12 guidance provided for scoring within these criteria, remain open to bias especially
13 from unforeseen circumstances that often fall beyond the process boundary, e.g.
17 nature and likelihood of these risks has become a priority for water producers,13 in
20 Lindley and Buchberger83 and Besner et al.,13 described earlier (see Asset
24
25 B. Reliability analysis
26
53
1 It is implicit in the planning, design and operation of water utilities that risk
5 component (e.g. risk of valve failure), process (e.g. risk of failure of treatment step) to
6 network (e.g. network reliability under drought conditions, see Network analysis)
7 level analysis. Regardless of focus, the aim is to identify the potential failures that
8 may occur in a system, their effects and their likelihood, thus aiding the identification
9 of critical components and processes where design and operational changes are
11 follows:151
12
18 sensitivity analysis
19
20 The National Health and Medical Research Council (NHMRC), the body
21 responsible for issuing drinking water guidelines to Australian water utilities, in their
24 the determination of critical control points whereupon risks can be monitored and
26 Water (Australia). The process begins with the division of the water system into four
54
1 discrete subsystems: catchment, treatment, distribution and customer premises.
2 Across each subsystem (e.g. catchment) the sources of risk to water quality (e.g.
3 native animals) and the associated hazards (e.g. bacteria, viruses) are identified and
4 plotted on a simple risk matrix; those risks deemed to be significant are evaluated
5 further for their critical control points. Assessors then identify the critical limits,
6 monitoring systems and corrective actions for each CCP. The application of HACCP
9 company has developed a greater understanding of water quality issues, refined and
11 These benefits stem from an increased knowledge and understanding of the water
12 supply system and an improved ability to identify potential risks to water supply /
13 quality.101 Beyond managing existing process control, HACCP may also be used to
14 assess and manage the risks from proposed operational changes, such as the
19 intentions of new or existing facilities in order to identify the hazards that may arise
21 examines a process (e.g. disinfection) subdivided into nodes, at each node, the team
22 applies guidewords (e.g. low) to process parameters (e.g. ozone levels) to identify
23 ways in which the process may deviate from its design intention, before evaluating the
55
1 response to regulatory requirements, identified the key areas of uncertainty (e.g.
3 formulated to clarify these uncertainties and to verify process conditions (e.g. check
4 pressure potential from the chlorine cylinder and the system response).
6 by the institutional capacity of organisations and the skill sets available at the
7 operational level. Risk analysis remains an expert discipline and many organisations
8 are more comfortable with the historic and proven implicit approach to risk
10 their analysis more explicit and using these tools for better decision-making,
11 identifying risk issues early rather than later, when their ability to respond may be
13 at the operational level. Targeted by a risk criticality scoring system, the analysis
14 systematically considers various components of the water supply system and their
16 makers are empowered to assess the costs and benefits in terms of risk reduction per
18 identified failure modes are traced to specific mechanical or electrical equipment, the
25 effectively using personnel with appropriate skills, experience and resources, they
26 provide operational management with a basis for improving process reliability and
56
1 identifying issues early. Ineffectively applied, they become little more than acronyms
3 treatment of the effect of identified risks at the system level. The importance and
4 complexity of this task has increased in recent years, due in part to the increased range
7 sufficient for this purpose. In the absence of such data, there is a requirement for the
8 formal modelling of risk consequences. There exist a range of techniques for this
9 task, including logic modelling (e.g. Demotier et al.;32 Cyna29), quantitative FMECA
10 (e.g. Cyna), and multiple barrier approaches to treatment reliability (e.g. Demotier et
13 developed and applied by the Compagnie Generale des Eaux (France) (Figure 7).
14 Following system definition and modelling (via reliability block diagrams), risks are
17 probability of the system to be found operative at a given time). Cyna describes how
19 Marne plant, arguing that its employment helped conceive a reliable system and
20 verified the adequacy of plant availability. The author concludes that reliability
22 design, and thus cost, to the level of reliability required, and, when associated with
24
25
26 5. CONCLUSION
57
1
3 paradigm, recognising the implicit approach performed over the last 150 years. With
5 that allow system vulnerabilities to be identified before failures occur are essential. In
6 many ways, however, the industry is discovering risk analysis afresh and there is a
7 learning curve to climb in terms of the capabilities and limitations of these tools and
8 techniques. The international water sector has helpfully restated its overarching goal
9 reminding us that even in the face of rationalisation and economic pressure, public
10 health protection8 is the principal business of the water industry. Risk analysis has a
11 part to play in focussing effort in the right places, but should not be treated as a
12 panacea or substitute for managing risk and neither allowed to dictate the outcome of
15 selection within the organisational context and legal framework. For large multi-
16 utilities, one can expect high developed business risk capabilities, whereas for smaller
17 and single utilities, an approach based on accepted codes and standards may be more
19 art as a reference for developing a risk analysis strategy that is fit for purpose.
20
21 ACKNOWLEDGEMENTS
22
23 This work has been funded by the American Water Works Association Research
26 organisations. The comments and views herein are the authors alone. Brian
58
1 MacGillivray is co-funded on a UK Engineering and Physical Sciences Research
4 REFERENCES
8 1992, 64-67
10 transport: the role of random spatial variability and sorption kinetics, Ground
20 http://www.adb.org/Documents/Handbooks/Water_Supply_Projects/default.asp)
24 http://www.adb.org/Documents/Conference/Water_Utilities/default.asp).
59
1 8. AWWA, EUREAU, WSAA and DVGW, Bonn Workshop 2001 Key principles
6 10. Barreix, A., Jenkins, G.P. and Marchesini, M., Case 1. Restructuring the water
8 http://130.15.74.22/publications/panama_water.pdf).
9 11. Beck, M.B., Applying systems analysis in managing the water environment:
11 12. Berry, M.A., Burmeister, E. and McElroy, M.B., Sorting out risks using known
13 13. Besner, M-C., Gauthier, V., Barbeau, B. and Millette, R., Understanding
15 14. Black, F. and Scholes, M., The pricing of options and corporate liabilities, J.
17 15. Boak, R.A. and Packman, M.J., A methodology for the assessment of risk of
20 16. Booth, R. and Rogers, J., Using GIS technology to manage infrastructure capital
22 17. Brammer, L.F. and Schulte, A.M., A joint US and UK approach to water supply
24 18. Brueck, T., Facing the workforce reality: making more out of less, J. AWWA,
60
1 19. Buckland, R. and Fraser, P., Political and regulatory risk in water utilities: Beta
4 20. Bunn, D.W., Dyner, U. and Larsen, E.R., Modelling latent market power across
5 gas and electricity markets, Syst. Dynam. Rev., 13, 271, 1997.
6 21. Burchett, J.F. and Tummala, V.M.R., The application of risk management
9 22. Clark, T., Jeffrey, P. and Stephenson, T., Complex agendas for new technology
11 23. Clemons, E.K., Using scenario analysis to manage the strategic risks of
13 24. Codex Alimentarius, Guidelines for the Application of the Hazard Analysis and
16 25. Cole, J.A., Hawker, P.J., Lawson, J.D. and Montgomery, H.A.C., Pollution risks
18 26. Colmer, G., Dunkley, M., Gray, K., Pugh, P. and Williamson, A., Estimating the
19 cost of new technology products, Intl. J. Technol. Manage., 17(7/8), 840, 1999.
20 27. Cowan, S., Competition in the water industry, Oxford Rev. of Econ. Policy,
22 28. Crowe, T.J., Fong, P.M., Bauman, T.A. and Zayas-Castro, J.L., Quantitative risk
25 29. Cyna, B., Reliability analyses of water treatment plants, Water Supply
61
1 30. Dabrowski, J.M., Peall, S.K.C., Van Niekerk, A., Reinecke, A.J., Day, J.A. and
5 31. Dean, A., Carlisle, Y. and Baden-Fuller, C., Punctuated and continuous change:
7 32. Demotier, S., Odeh, K., Schon, W., Charles, P., Fotoohi, F. and Allioux, J-F.,
9 http://www.hds.utc.fr/~tdenoeux/sime/publis/esrel2002.pdf.).
10 33. Denton, D.K., Keeping employees: the Federal Express approach, S.A.M. Adv.
12 34. Dewettinck, T.,Van Houtte, E., Geenens, D., Van Gege, K. and Verstraete, W.,
13 HACCP (Hazard Analysis and Critical Control Points) to guarantee safe water
14 reuse and drinking water production a case study, Water Science and
16 35. Dickerson, A.P., Gibson, H.D. and Tsakalotos, E., Is attack the best form of
21 37. Doyle G. and Grabinsky, M., Applying GIS to a water main corrosion study. J.
23 38. Drohan, W.M. and OConnor, K., Competitor analysis can give your association
25 39. Egerton, A.J., FMEA water flowsheet, Lectures in risk management for water
62
1 40. Egerton, A.J., Minimising the risk of contamination in water treatment, Lecture
4 www.egertonconsulting.co.uk).
5 41. Eisenberg, D., Soller, J., Sakaji, R. and Olivieri, A., A methodology to evaluate
6 water and wastewater treatment plant reliability, Wat. Sci. Technol., 43(10), 91,
7 2001.
8 42. Elias, D., Allocating capital risk: Who should pay?, Wat. Engg. Manage.,
10 43. Fedra, K., Integrated risk assessment and management: overview and state of
12 44. Feijtel, T., Boeije, G., Matthies, M., Young, A., Morris, G., Gandolfi, C.,
13 Hansen, B., Fox, K., Matthijs, E., Koch, V., Schroder, R., Cassani, G.,
17 45. Fewtrell L. and Bartram J., Water quality: guidelines, standards and health
20 46. Fisher, M.L., What is the right supply chain for your product?, Harvard
22 47. Fitzpatrick, E.W., Assessing the worth of new technology, J. Amer. Soc. of CLU
24 48. Foran, J., Brosnan, T., Connor, M., Delfino, K., Depinto, J., Dickson, K.,
25 Humphrey, H., Novotny, V., Smith, R., Sobsey, M. and Stehman, S., A
63
1 framework for comprehensive, integrated, watershed monitoring in New York
3 49. Foster, J.A. and McDonald, A.T., Assessing pollution risks to water supply
6 50. Foster, M., Fell, R. and Spannagle, M., A method for assessing the relative
8 1025, 2000.
9 51. Fuest, S., Berlekamp, J., Klein, M. and Matthies, M., Risk hazard mapping of
12 52. Gale, P., Developments in microbiological risk assessment for drinking water, J.
14 53. Gray, R.S. and Powell, E., Risk management of major aqueducts, J. IWEM, 2,
15 594, 1988.
18 55. Grimsey, D. and Lewis, M.K., Evaluating the risks of public private partnerships
21 management strategies for the proposed Isikli Reservoir, Wat. Sci. Technol.,
23 57. Haas, C.N. and Trussell, R.R., Frameworks for assessing reliability of multiple,
24 independent barriers in potable water reuse, Wat. Sci. Technol., 38(6), 1, 1998.
25 58. Haas, C.N., Progress and data gaps in quantitative microbial risk assessment,
64
1 59. Haas, C.N., Rose, J.B., Gerba, C.P., Quantitative Microbial Risk Assessment,
3 60. Hagevik, S., What keeps em? J. Environ. Hlth., 63(6), 50, 2001.
4 61. Halfacree, S., Pollution risk management, Issues in Environ. Sci. Technol., 9, 37,
5 1998.
6 62. Hartmann, G.C. and Lakatos, A.I., Assessing technology risk a case study,
10 64. Hellier, K., Hazard analysis and critical control points for water supplies, Proc.
13 65. Hern, R., Competition and access pricing in the UK water industry, Utilities
15 66. Hertz, D.B., Risk analysis in capital investment, Harvard Business Review,
17 67. Hrudey, S.E. and Leiss, W., Risk management and precaution insights on the
19 68. Hrudey, S.E., Limits to Science for Assessing and Managing Environmental
20 Health Risks, Science, Truth and Justice, Canadian Institute for the
23 69. Hunter, P.R., Waite, M., Ronchi, E., Drinking Water and Infectious Disease:
65
1 70. IMPRESS Management, IMPRESS Guidance, Guidance for the analysis of
2 pressures and impacts in accordance with the water framework directive, Draft
4 71. Jinno, K., Zongxue, X, Kawamura, A. and Tajiri, K., Risk assessment of a water
5 supply system during drought, Wat. Resource Develop., 11(2), 185, 1995.
6 72. Kaufman, M.M. and Wurtz, M., Small system maintenance management using
8 73. Kent, N.A., Treece, D. and Davies, M., Asset management through risk
11 prepared for the challenges ahead?, 14th November, 2003, London, UK.
12 74. Khan, F.I. and Abbasi, S.A., Towards automation of HAZOP with a new tool
14 75. Kilpatrick, K. and Lapsley, I., Trust us Regulators and the governance of
16 76. Laing, R.D., Report of the Commission of Inquiry into matters relating to the
19 (website: http://www.northbattlefordwaterinquiry.ca.)
20 77. Lantzy, R.J., Cichy, P.T. and Leitzinger, A.H., Use of geographical maps to
24 strategic and regulatory risk, J. Oper. Res. Soc., 50, 337, 1999.
25 79. Latiffe, R., Probabilistic risk analysis of large dams: its value and limits, Intl.
66
1 80. Latinopoulos, P., Mylopolous, N. and Mylopoulos, Y., Risk-based decision
2 analysis in the design of water supply projects, Wat. Resources Manage., 11,
3 263, 1997.
4 81. Li, S.G. and McLaughlin, D., A nonstationary spectral method for solving
7 82. Lifton, G. and Smeaton, P., Asset risk management in Scottish Water, Proc. of
9 who is best prepared for the challenges ahead?, 14th November, 2003, London,
10 UK.
13 84. Liou, T-S. and Yeh, H-D., Conditional expectation for evaluation of risk
17 HMSO, 1986.
19 management model for outsourcing, Supply Chain Manage., 4(4), 176, 1999.
22 88. Lytton, L., Howe, S., Sage, R. and Greenaway, P., Groundwater abstraction
67
1 90. Masago, Y., Katayama, H., Hashimoto, A., Hirata, T. and Ohgaki, S.,
4 91. Maxwell, S., Where are we headed? Ten key trends and developments in the
6 92. McGaughey Jr., R.E., Synder, C.A. and Carr, H.H., Implementing information
9 93. McIntyre, N.R., Wagener, T., Wheater, H.S. and Chapra, S.C., Risk-based
12 94. McNally, J.S., Retaining valued employees, Pennsylvania CPA J., 71(4), 24,
13 2001.
14 95. Medema, G.J., Hoogenboezem, W., van der Veer, A.J., Ketelaars, H.A.M.,
17 2003.
18 96. Merabtene, T., Kawamura, A., Jinno, K. and Olsson, J., Risk assessment for
21 97. Modigliani and Miller, The cost of capital, corporation finance and the theory of
23 98. Morana, C. and Sawkins, J.W., Regulatory uncertainty and share price volatility:
24 the English and Welsh water industrys periodic price review, J. Regul. Econ.,
68
1 99. Mosca, R., Schenone, M. and Bonello, A.F., Risk analysis in the evaluation of
4 100. Mowday, R.T., Strategies for adapting to high rates of employee turnover,
6 101. Mullenger, J., Ryan, G., and Hearn, J., A water authoritys experience with
12 104. Newman, W.R., Hanna, M. and Maffei, M.J., Dealing with the uncertainties of
17 106. NRDC, What's on tap? Grading drinking water in US cities, Natural Resources
19 http://www2.nrdc.org/water/drinking/uscities/pdf.).
20 107. NRWA, Security vulnerability self-assessment guide for small drinking water
22 http://www.health.ri.gov/environment/dwq/security/vulnerability.htm).
23 108. OConnor D.R., Report of the Walkerton Inquiry. Part 1. The events of May
24 2000 and related issues, The Attorney General of Ontario, Toronto, Canada,
25 2002.
69
1 109. OReilly, C.A., Chatman, J.A. and Caldwell, D., People and organizational
4 110. Ofwat, Access codes for common carriage: guidance, Ofwat, 2002. (website:
5 http://www.ofwat.gov.uk).
6 111. Osowski, S.L., Swick Jr, J.D., Carney, G.R., Pena, H.B., Danielson, J.E. and
9 159, 1999.
10 112. Ostfeld, A., Kogan, D. and Shamir, U., Reliability simulation of water
12 113. Ostfeld, A., Reliability analysis of regional water distribution systems, Urban
14 114. Parker, D., Performance, risk and strategy in privatised, regulated industries,
16 115. Parker, D., Reducing regulatory risk: the case for a new regulatory contract with
18 116. Parmelee, M.A., Trends in outsourcing for water utilities, J. AWWA, 94(9), 54,
19 2002.
20 117. Parr, N.M. and Cullen, N., Risk management and reservoir maintenance, J.
22 118. Passarella, G., Vurro, M., DAgostino, V., Guiliano, G. and Barcelona, M.J., A
25 119. Payment, P., Richardson, L., Siemiatycki, J., Dewar, R., Edwardes, M. and
26 Franco, E., A randomized trial to evaluate the risk of gastrointestinal disease due
70
1 to consumption of drinking water meeting currently accepted microbial
3 120. Phelps, R., Chan, C., Kapsalis, S.C., Does scenario planning affect
5 121. Pollard, S.J.T., Strutt, J.E., MacGillivray, B.H., Hamilton, P.D. and Hrudey,
6 S.E., Risk analysis and management in the water utility sector a review of
7 drivers, tools and techniques, Trans. IChemE, Part B: Proc. Safety Environ.
9 122. Pollard, S.J.T., Yearsley, R., Reynard, N., Meadowcroft, I.C., Duarte-Davidson,
11 assessment in the United Kingdom, Environ. Sci. Technol., 36(4): 530, 2002.
12 123. Portielje, R., Hvitved-Jacobsen, T. and Schaarup-Jensen, K., Risk analysis using
16 handle risk and uncertainty, The Strategy Unit, London, 2002. (website:
17 http://www.number-10.gov.uk/SU/RISK/risk/home.html.).
19 corporate strategy: outsourcing motives and risks, Eur. Manage. J., 21(5), 647,
20 2003.
21 126. Radovanovic, R. and Marlin, A., Critical water mains the need for risk
23 127. Rajani, B.B. and Kleiner, Y., WARP Water Mains Renewal Planner, National
25 cnrc.gc.ca/fulltext/nrcc44680/)
71
1 128. Ranasinghe, M., Private sector participation in infrastructure projects: a
3 1999.
6 130. Remenyi, D. and Heafield, A., Business process re-engineering: some aspects of
7 how to evaluate and manage the risk exposure, Intl. J. Proj. Manage., 14(6),
8 349, 1996.
9 131. Rizak, S., Cunliffe, D., Sinclair, M., Vulcano, R., Howard, J., Hrudey, S. and
10 Callan, P., Drinking water quality management: a holistic approach, Wat. Sci.
12 132. Rogers, L.L., Dowla, F.U. and Johnson, V.M., Optimal field-scale groundwater
13 remediation using neural networks and the genetic algorithm, Environ. Sci.
15 133. Rothschild, W.E., Competitor analysis: the missing link in strategy, The
19 135. Saal, D.S. and Parker, D., Productivity and price performance in the privatized
20 water and sewerage companies of England and Wales, Journal of Regul. Econ.,
22 136. Sawkins, J.W., The development of competition in the English and Welsh water
24 137. Sawkins, J.W., Yardstick competition in the English and Welsh water industry:
72
1 138. Seidenstat, P., Emerging competition in water and wastewater industries.
3 www.ucowr.siu.edu/updates/pdf/V117_A2.pdf).
4 139. Seker, D.Z., Kadbasli, S. and Rudvan, B., Risk assessment of a dam-break using
6 140. Sharpe, W.F., Capital asset prices: a theory of market equilibrium under
10 142. Shimoda, T.A., Emergency preparedness and response, J. AWWA, 86, 84, 1994.
11 143. Shleifer, A., A theory of yardstick competition, Rand J. Econ., 16(3), 319, 1985.
12 144. Sigler, K.J., Challenges of employee retention, Manage. Res. News, 22(10), 1,
13 1999.
14 145. Singh, H. and Harianto, F., Management-board relationships, takeover risk, and
16 146. Sivertun, A. and Prange, L., Non-point source critical area analysis in the
17 Gissel watershed using GIS, Environ. Modelling and Software, 18, 887, 2003.
18 147. SNL news release, Sandia develops program to assess water infrastructure
20 http://www.sandia.gov/media/NewsRel/NR2001/watinfr.htm).
21 148. Songer, A.D., Diekmann, J. and Pecsok, R.S., Risk analysis for revenue
23 149. Stahl, G.R. and Elliott, J.C., New generation computer models for risk-based
73
1 150. Stevens, I.M. and Lloyd, R.I., Water resources: Sustainable optimisation and
4 151. Strutt, J.E., Reliability data and sources, Lectures in risk management for water
6 152. Sweeney, M.W., Geographic information systems, Wat. Environ. Res., 71(5),
7 551, 1999.
8 153. Ta, C.T., A probability model for burst risk studies of water mains, Wat. Sci.
13 155. Tebutt, R.J., Guy, J.A., Gochin, R.J. and Lester, J.N., Requirements of a capital
14 investment appraisal technique for the water industry in England and Wales,
16 156. Teunis, P.F.M., Medema, G.J, Kruidenier, L. and Havelaar, A.H., Assessment of
19 157. Thomas, H., Risk analysis and the formulation of acquisition / diversification
21 158. Thompson, R., Secrets to keeping good employees, Business Credit, 102(1), 68,
22 2000.
25 160. UKWIR, Managing microbial and chemical risks from source to tap, Report
74
1 161. Umweltbundesamt, Water safety conference abstracts, Scholl, O. and Chorus, I.,
7 asset management has potential to help utilities better identify needs and plan
9 http://www.gao.gov/new.items/d04461.pdf).
10 164. Verro, R., Calliera, M., Maffioli, G., Auteri, D., Sala, S., Finizio, A. and Vighi,
13 165. Wang, J., Yuan, X., Zhang, Y. and Zhang, Y., A reliability and risk analysis
15 2003.
17 2003.
18 167. WHO, Guidelines for drinking water quality, 3rd Edition (draft), WHO, 2003.
19 (website: http://www.who.int/water_sanitation_health/dwq/guidelines3rd/en/).
22 169. Wickham, J.D. and Wade, T.G., Watershed level risk assessment of nitrogen
23 and phosphorus export, Computers and Electronics in Agriculture, 37, 15, 2002.
24 170. Wildemann, H., Justification on strategic planning for new technologies, Human
75
1 171. Wingle, W.L., Poeter, E.P. and McKenna, S.A., UNCERT: geostatistics,
4 172. World Bank, Toolkits for private participation in water and sanitation, World
6 http://www.worldbank.org/html/fpd/water/wstoolkits/index.html).
9 174. Zongxue, X., Jinno, K., Kawamura, A., Takesaki, S. and Ito, K., Performance
10 risk analysis for Fukuoka water supply system, Wat. Resources Manage., 12, 13,
11 1998.
12 175. Zoppou, C., Review of urban storm water models, Environ. Modelling and
14 176. Zsidisin, G.A., Panelli, A. and Upton, R., Purchasing organization involvement
17
18
19
20
21
22
23
24
25
26
27
28
76
1
8 Figure Captions
9 Figure 1. Decision Framework for the Offshore Oil Industry (UK Oil Operators
11 Figure 2. The risk hierarchy (adapted from Prime Ministers Strategy Unit124).
17
18
19
20
21
22
23
24
25
26
77
1
78
SIGNIFICANC E TO DEC ISION
MEANS O F CALIBRATIO N MAKING PROC ESS DEC ISION CONTEXT TYPE
Internal Stakeholder
Cons ultation Very novel or challenging
Company Strong stakeholder views
values Significant risk trade-offs / trans fers
C Large uncertainties
External Stakeholder Perceived lowering of safety stds.
Cons ultation Societal
values
3 Figure 1. Decision Framework for the Offshore Oil Industry (UK Oil Operators
10
79
1
2
Regulatory risk
Competition risk
Strategic decisions Business process re-engineering
Decisions required
for implementation Compliance risk
Operational Reliability analysis
5 Figure 2. The risk hierarchy (adapted from Prime Ministers Strategy Unit124).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
81
1 Technology risk
No physical Very high
2 analysis completed
3
Near physical High
4 Not solved limit
before
5
Physical analysis
6 completed
7
Not near
8 physical limit Medium
10 Technology
problem
11
12 No physical High
analysis completed
13
14
Solution not
accessible to High
15
utility
16 Solved
before
17 Near
physical
18 limit
Solution
19
accessible to
utility Medium
20
21 Physical
analysis
22 completed
Solution not
accessible to Medium
23
utility
24
Not near
25
physical
limit
26
Solution
27 accessible to Low
28 utility
29
30 Figure 3. Technology risk algorithm (Hartmann and Lakatos62 with permission).
82
1
OR
Gate
AND AND
Gate Gate
3 Tr = resource turbidity
4 The probability of the top-event may be calculated if the probabilities of the sub-events are known or
5 estimable.
7 Figure 4. Illustrative fault tree for turbidity non-compliance (after Demotier et al.31).
10
11
12
13
14
15
16
17
18
19
83
1 a) Series diagram here, the system / process is working if A1 and A2 are
2 working
3
4
5 Input A1 A2 Output
6
7
8
9 b) Parallel diagram here, the system / process is working if A1 or A2 are
10 working
11
12
13
14 A1
15
16
17 Output
18 Input
19
20
21
A2
22
23
24
25
26 c) Redundancy diagram here, the system / process is working if at least r
27 elements among n are working
28
29
30
A1
31
32
33
34
35 A2 r/n Output
Input
36
37
38
39
40 An
41
42
43
44
45
46 Figure 5. Reliability block diagrams (after Cyna29).
47
48
49
84
1
2
3
4
5 Figure 6. Risk-based workforce planning (after Pollard et al.122).
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
85
1
2
3
Description of the system
4
Establish Availability goals
5
6
7
8
Create system model
9
10
11
12
13
14 Research of reliability data
15
16
17
18
19 Identification of failure modes
20 (HAZOP)
21
22
23
24 Quantification of failures (FMECA)
25
26
27
28
29
Availability computation
30
31
32
33
34
35
Comparison of computed
36
availability with goal
37
38
39
40
41 Figure 7. Methodology for reliability analysis of a water treatment plant (after
42 Cyna29).
86
Table 1. Strategic risk portfolio.
87
Table 2. BPR risk matrix (after Remenyi and Heafield130).
FACTORS FACTORS
Novelty of technology
Outsourcing utilisation
Skills base
88
Table 3. Programme level risk portfolio.
GIS risk simulation Quantified risk mapping over space and time Dabrowski et al.,30 Verro et al.,164 Feijtel et al.44
Network Network reliability a) Assess susceptibility to supply-demand scenarios; b) aid Stevens and Lloyd,150Lifton and Smeaton,82 Wang et al.,165
analysis modelling development of supply strategies and policies; c) assist Merabtene et al.,96 Ostfeld,113 Stahl and Elliot,149 Zongxue et
design of distribution networks; and d) inform the need for al.,174 Andreu et al.,3 Jinno et al.71
capital expenditure.
Vulnerability RAM-W To assess system vulnerabilities and develop measures to SNL news release.147
assessment reduce risks of attack.
89
Table 4. Component FMEA for chlorine cylinder and outlet valve (Egerton39 with permission of Egerton Consulting Ltd.).
Failure mode Failure effect on process Failure effect on system Methods of Detection Comments
Fail to open / Loss of adequate chlorination Non-potable water will leave plant Changeover should detect loss System failure would
Reduced output / of supply require combination of
No output loss of flow and failure of
changeover
Fail to close None changeover should None None
transfer to standby cylinders
Excess output Excess chlorination Possible taste and odour Changeover should detect
complaints. No serious excess chlorine flow
consequences
Outside specification Outside specification (wrong Non-potable water will leave plant. QA checks on delivery. Low
(wrong or or contaminated gas) POTENTIAL FOR MAJOR chlorine residual readings and
contaminated gas) SAFETY HAZARD alarm
90
Table 5. Operational level risk portfolio.
91